|
|
| Author |
Message |
Carl R. science forum beginner
Joined: 01 May 2005
Posts: 38
|
Posted: Wed May 04, 2005 6:21 am Post subject:
Simplicial approximation help
|
|
|
Hi, can you please explain the following solution to the
problem:
This is an example to illustrate that simplicial approximation
may not exist.
Let f: [0,1]->[0,1] be the map given by x->x^2.
Let K=L be the simplicial complex with vertices 0,1/2,1
and edges <0,1/2> and <1/2, 1>.
The solution consists in showing that we can't find
an admissible vertex map phi: V(K) (set of vertices of K)
to V(L) (set of vertices of L) such that the induced
continuous map |phi|: |K| -> |L| where |K| denotes the
underlying space of K.
So clearly f(0)=0^2=0 and hence carrier(f(0))=carrier(0)
=<0>. And hence since |phi| is an approximation of f
then |phi(0)| belongs to carrier(f(0))=<0> and hence
|phi(0)|=0.
Ok so far this is clear. The next part is the one I don't get.
Then the author says:
Since |phi(0)|=0 this implies phi(0)=0. Why is this???
I understand that since <0> is a simplex of K then by
definition of admissible vertex map <phi(0)> must be mapped to
a vertex of L, and the only vertices of L are
0,1/2 and 1, hence phi(0)=0 or 1/2 or 1. But why the author
concludes phi(0)=0 ?
Then it says , f(1)=1^2 and hence phi(1)=1.
Again I don't get why necessarily this implies phi(1)=1.
Then:
Since phi is an admissible vertex map we must have phi(1/2)=1/2.
How do you know that 1/2 must get mapped to 1/2 and not other point?
Then it says for x=2/3 , f(x)=4/9 with carrier <0,1/2>.
So the carrier condition fails for this point since
|phi|(2/3)= 2/3 which is not in <0,1/2>.
Again my question why is |phi|(2/3) = 2/3?
Thanks in advance. |
|
| Back to top |
|
 |
John H Palmieri science forum beginner
Joined: 04 May 2005
Posts: 13
|
Posted: Wed May 04, 2005 1:59 pm Post subject:
Re: Simplicial approximation help
|
|
|
On May 04 2005, "Carl R." <solrac140@hotmail.com> wrote:
| Quote: | Hi, can you please explain the following solution to the
problem:
|
Overall, I think you don't understand what |phi| is. Check the
definition again. I have more specific comments below.
| Quote: | This is an example to illustrate that simplicial approximation
may not exist.
Let f: [0,1]->[0,1] be the map given by x->x^2.
Let K=L be the simplicial complex with vertices 0,1/2,1
and edges <0,1/2> and <1/2, 1>.
The solution consists in showing that we can't find
an admissible vertex map phi: V(K) (set of vertices of K)
to V(L) (set of vertices of L) such that the induced
continuous map |phi|: |K| -> |L| where |K| denotes the
underlying space of K.
So clearly f(0)=0^2=0 and hence carrier(f(0))=carrier(0)
=<0>. And hence since |phi| is an approximation of f
then |phi(0)| belongs to carrier(f(0))=<0> and hence
|phi(0)|=0.
Ok so far this is clear. The next part is the one I don't get.
Then the author says:
Since |phi(0)|=0 this implies phi(0)=0. Why is this???
I understand that since <0> is a simplex of K then by
definition of admissible vertex map <phi(0)> must be mapped to
a vertex of L, and the only vertices of L are
0,1/2 and 1, hence phi(0)=0 or 1/2 or 1. But why the author
concludes phi(0)=0 ?
|
If phi(0)=a, then |phi|(0)=a, by definition of |phi|. Therefore
since |phi|(0)=0, phi(0) can't be 1/2 or 1.
| Quote: | Then it says , f(1)=1^2 and hence phi(1)=1.
Again I don't get why necessarily this implies phi(1)=1.
|
Same reason.
| Quote: | Then:
Since phi is an admissible vertex map we must have phi(1/2)=1/2.
How do you know that 1/2 must get mapped to 1/2 and not other point?
|
Where else can 1/2 go? If phi(1/2)=0, for example, then where would
phi take the edge from 1/2 to 1?
| Quote: | Then it says for x=2/3 , f(x)=4/9 with carrier <0,1/2>.
So the carrier condition fails for this point since
|phi|(2/3)= 2/3 which is not in <0,1/2>.
Again my question why is |phi|(2/3) = 2/3?
|
If phi(1/2)=a and phi(1)=b, then |phi| takes the edge from 1/2 to 1,
linearly, to the edge from a to b. In this case a=1/2 and b=1, so
|phi| is the identity map on that edge.
| Quote: | Thanks in advance.
|
--
J. H. Palmieri
Associate Professor of Mathematics University of Washington
Box 354350, Seattle, WA 98195-4350 palmieri@math.washington.edu
http://www.math.washington.edu/~palmieri/ |
|
| Back to top |
|
 |
Google
|
|
| Back to top |
|
 |
|
|
The time now is Fri Jan 09, 2009 8:51 pm | All times are GMT
|
|
Loans | Loans | Homeowner Loans | Loan | Loans
|
|
Copyright © 2004-2005 DeniX Solutions SRL
|
|
Other DeniX Solutions sites:
Electronics forum |
Medicine forum |
Unix/Linux blog |
Unix/Linux documentation |
Unix/Linux forums
|
Powered by phpBB © 2001, 2005 phpBB Group
|
|